ARTICLE

DOI: 10.1038/s41467-018-06836-4 OPEN MicroRNA-27a controls the intracellular survival of Mycobacterium tuberculosis by regulating calcium- associated autophagy

Feng Liu1, Jianxia Chen1,2, Peng Wang3,4, Haohao Li3, Yilong Zhou3, Haipeng Liu 1,2, Zhonghua Liu1, Ruijuan Zheng1, Lin Wang3, Hua Yang1, Zhenling Cui1, Fei Wang3, Xiaochen Huang1, Jie Wang1, Wei Sha4, Heping Xiao4 & Baoxue Ge1,2,3 1234567890():,; Tuberculosis (TB) caused by Mycobacterium tuberculosis (Mtb) kills millions every year, and there is urgent need to develop novel anti-TB agents due to the fast-growing of drug-resistant TB. Although autophagy regulates the intracellular survival of Mtb, the role of calcium (Ca2+) signaling in modulating autophagy during Mtb infection remains largely unknown. Here, we show that microRNA miR-27a is abundantly expressed in active TB patients, Mtb-infected mice and macrophages. The target of miR-27a is the ER-located Ca2+ transporter CAC- NA2D3. Targeting of this transporter leads to the downregulation of Ca2+ signaling, thus inhibiting autophagosome formation and promoting the intracellular survival of Mtb. Mice lacking of miR-27a and mice treated with an antagomir to miR-27a are more resistant to Mtb infection. Our findings reveal a strategy for Mtb to increase intracellular survival by manip- ulating the Ca2+-associated autophagy, and may also support the development of host- directed anti-TB therapeutic approaches.

1 Shanghai Key Lab of Tuberculosis, Shanghai Pulmonary Hospital, Tongji University School of Medicine, Shanghai 200433, China. 2 Clinical Translation Research Center, Shanghai Pulmonary Hospital, Tongji University School of Medicine, Shanghai 200433, China. 3 Department of Microbiology and Immunology, Tongji University School of Medicine, Shanghai 200092, China. 4 Department of TB, Shanghai Pulmonary Hospital, Tongji University School of Medicine, Shanghai 200433, China. These authors contributed equally: Feng Liu, Jianxia Chen, Peng Wang. Correspondence and requests for materials should be addressed to B.G. (email: [email protected])

NATURE COMMUNICATIONS | (2018) 9:4295 | DOI: 10.1038/s41467-018-06836-4 | www.nature.com/naturecommunications 1 ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-06836-4

ycobaterium tuberculosis (Mtb) remains the leading autophagy27. Upregulation of miRNA-125a in Mtb-infected bacterial cause of death in humans throughout the macrophages targets UV radiation resistance-associated gene M 1 28 world . In 2015, Mtb infection was responsible for 10.4 (UVRAG) to prevent autophagy . More recently, Kim et al. million new cases and 1.4 million death2. However, only ∼10% of found that miR-144* is induced by Mtb to target DNA damage infected individuals develop active tuberculosis (TB), while the regulated autophagy modulator 2 (DRAM2), resulting in sup- remaining 90% of cases exhibit latent infection, indicating a cri- pression of autophagy29. However, our understanding of the role tical role for host immunity in limiting Mtb infection3. of miRNAs in modulating autophagic signaling pathways during Autophagy is a cellular process that can form double Mtb infection and their clinical relevance is still limited. membrane-layered vesicles to deliver macromolecules, organelles, In this study, we found that miR-27a is abundantly expressed or intracellular pathogens for lysosomal degradation4,5. A num- in active TB patients, Mtb-infected animals and infected cells. ber of molecules including autophagy-related ATG family pro- Induction of miR-27a was found to directly target the Ca2+ teins orchestrate signaling events that regulate autophagy flux transporter Cacna2d3, and down-regulate ER Ca2+ signaling to including autophagosome initiation, elongation, maturation, and inhibit autophagy, thus promoting the intracellular survival fusion with lysosomes4–6. As a secondary messenger, intracellular of Mtb. Ca2+ acts as a major regulator of autophagy. Intracellular Ca2+ is actively pumped out of the plasma membrane from the cytosol or Results is sequestered into multiple organelles including the endoplasmic Induction of miR-27a expression by Mtb. We evaluated miRNA reticulum (ER), mitochondria, and lysosomes7. Upon stimulation, + expression profiles from peripheral blood mononuclear cells entry of Ca2 into the cytoplasm activates CaMKKβ,a + (PBMCs) of the patients with active pulmonary TB, the lungs of Ca2 /calmodulin-dependent serine/threonine kinase8. Once Mtb strain H37Rv-infected C57/BL6 mice, and H37Rv-infected activated, CaMKKβ phosphorylates AMPK on Thr172, leading to murine primary peritoneal macrophages. Using miRNA deep its activation8. Activated AMPK has been shown to inhibit sequencing, we found that miR-27a, a microRNA mainly impli- mTORC1, which phosphorylates the serine/threonine protein – cated in tumor cells30 32, was commonly upregulated in all kinase ULK1 (unc-51-like autophagy activating kinase 1, the samples compared to their representative controls (Fig. 1a–c). As mammalian orthologue of yeast ATG1) on Ser757, preventing the miR-27a’s functional role in regulation of Mtb infection remains association of ULK1 with AMPK9. Phosphorylation of AMPK on uncharacterized, we choose it for our further study. The upre- Thr172 by CaMKKβ thus enables AMPK to interact with and gulation of miR-27a was validated by quantitative real time PCR phosphorylate ULK1 on Ser555. AMPK-phosphorylated ULK1 (RT-PCR) in PBMCs from human TB samples, the lungs of forms a complex with FIP200 (200 kDa focal adhesion kinase H37Rv-infected C57/BL6 mice, and H37Rv-infected macrophages family-interacting protein) and ATG13 to induce autophagy by (Supplementary Fig. 1a-c). However, Mtb-induced miR-27a phosphorylating Beclin-1 and activating VPS34 lipid kinase10. + expression was markedly attenuated in Tlr2-/- macrophages but Recently, it has been reported that lysosomal Ca2 released not in Tlr4-/- macrophages (Supplementary Fig. 1d). Consistently, through mucolipin 1 (MCOLN1) activates the phosphatase cal- among the TB component examined, only stimulation with the cineurin, which dephosphorylates TFEB, thus promoting its TLR2 ligand peptidoglycan (PGN), but neither of the TLR4 nuclear translocation to induce lysosomal biogenesis and autop- ligand lipopolysaccharide (LPS) nor the Gram-negative bacterium hagy11. However, whether and how non-lysosomal calcium sig- Escherichia coli, TDM, ESAT-6 and ManLAM, induced the naling can also regulate autophagy still remains unclear. expression of miR-27a (Supplementary Fig. 1e). To further Mtb, as a successful intracellular pathogen, that can survive and investigate the signaling pathway involved in the induction of persist in macrophage by blocking the fusion of phagosome with miR-27a, we infected macrophages with Mtb in the presence of a lysosome12. Recent work demonstrated that autophagy played an selective p38 inhibitor (SB203580), MEK inhibitor (PD98059), or important role in the containment of intracellular Mtb by tar- NF-κB inhibitor (pyrrolidine dithiolar-bamate, PDTC). Selective geting Mtb for lysosome degradation, bypassing Mtb-mediated inhibition of the MEK-ERK pathway by PD98059 or NF-κB blockade of phagosome maturation13. Stimulation of autophagy pathway by PDTC markedly reduced Mtb-induced miR-27a by autophagy-inducing agents or interferon-γ (IFN-γ) increased expression in macrophages (Supplementary Fig. 1f). These results the delivery of Mtb to autophagosomes and lysosome-mediated – suggested that Mtb infection may induce miR-27a expression via bacterial killing13 16. However, in resting cells, the majority of the activation of ERK MAPK and NF-κB pathways. Mtb is not found to be associated with an autophagosome com- ponent such as LC3. Moreover, when unstimulated, autophagy is only a minor contributor to the clearance of Mtb in macrophages miR-27a promotes intracellular survival of Mtb. To examine the and in vivo in mice14,17–19, which may be due to the ability of functional role of miR-27a during Mtb infection, we analyzed its Mtb to partly disable or manipulate autophagy through multiple effect on the intracellular survival of Mtb by colony-forming unit mediators, such as secretory proteins ESAT-6/CFP-10 and (CFU) assay. Primary murine peritoneal macrophages were PE_PGRS4720,21, or the cell wall component lipoarbinomannan transiently transfected with a miR-27a mimic or inhibitor and (LAM)22. were then infected with H37Rv at an MOI of 5 for 24 h. As MicroRNAs (miRNAs) are a family of small noncoding RNA transfection of miR-27a mimic or inhibitor had no effect on molecules ∼21–25 nucleotides long that bind to the 3′ untrans- phagocytosis of Mtb by macrophages, treatment with the miR-27a lated region (3′ UTR) of target mRNAs and reduce protein mimic significantly promoted the intracellular survival of Mtb in expression by blocking mRNA translation and/or by promoting macrophages, whereas the miR-27a inhibitor reduced the intra- mRNA degradation23,24. Recent studies underscore the critical cellular survival of Mtb (Fig. 1d, e and Supplementary Fig. 2a). role of miRNAs in the evasion of Mtb from autophagic clearance LDH assay were also performed and the result showed that miR- of macrophages25. It has been shown that Mtb upregulates miR- 27a inhibitor or mimic did not affect the cytotoxicity of macro- 33 to suppress several key autophagy effector molecules, such as phages induced by Mtb infection (Supplementary Fig. 2b). ATG5, ATG12, LAMP1, LC3B, AMPK, and FOXO3 in macro- To better characterize the function of miR-27a, we generated phages to inhibit autophagy26. Kumar et al. have reported that miR-27a-/- mice by using the CRISPR-Cas9 genome editing Mtb downregulates the expression of miR-17 to enhance its target method33 and qPCR results showed that miR-27a’s expression proteins Mcl-1, which in turn interacts with BECLIN-1 to inhibit but not the other members of the miR-23/24/27 cluster, was

2 NATURE COMMUNICATIONS | (2018) 9:4295 | DOI: 10.1038/s41467-018-06836-4 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-06836-4 ARTICLE

abcColor key and histogram –3.00 –2.00 –1.00 0.00 1.00 2.00 3.00 8 –2.58 –1.72 –0.86 0.00 0.86 1.72 2.58 0 Count 0 1234 mmu-miR-154-5p mmu-miR-26b-3p mmu-miR-150-5p mmu-miR-1964-3p mmu-miR-18a-3p mmu-miR-1981-5p hsa-miR-582-3p mmu-miR-1249-3p mmu-miR-125a-3p mmu-miR-181c-5p mmu-miR-342-5p hsa-miR-1249-3p mmu-miR-106a-5p mmu-miR-24-3p mmu-miR-222-5p mmu-miR-340-3p hsa-miR-4286 mmu-miR-20b-5p mmu-miR-28c mmu-miR-23b-3p mmu-miR-331-5p mmu-miR-338-5p hsa-miR-504-5p mmu-miR-872-3p mmu-miR-140-5p mmu-miR-125b-5p mmu-miR-novel-chrX mmu-miR-28a-5p hsa-miR-543 mmu-miR-31-5p mmu-miR-339-5p mmu-miR-381-3p mmu-miR-27a-3p hsa-miR-24-2-5p mmu-miR-27a-3p mmu-miR-337-5p mmu-miR-181a-1-3p mmu-miR-1198-5p hsa-miR-450a-1-3p mmu-miR-22-3p mmu-miR-27a-5p mmu-miR-144-3p mmu-miR-340-5p hsa-miR-27a-3p mmu-miR-3068-3p mmu-miR-335-3p mmu-miR-132-3p mmu-miR-125a-5p hsa-miR-130a-3p mmu-miR-5079-5p mmu-miR-19a-5p mmu-miR-212-3p mmu-miR-1249-3p mmu-miR-1249-5p hsa-miR-4659a-3p mmu-miR-154-3p mmu-miR-136-3p mmu-miR-125b-2-3p mmu-miR-331-3p mmu-miR-296-5p hsa-miR-2277-5p mmu-miR-376b-3p mmu-miR-34b-3p 0 h 4 h 1# 2# 3# 1# 2# 3# 0 day 28 days Mφ, Healthy TB C57B/L6 H37Rv (MOl=5) control patients H37Rv (CFU=200)

defNC g miR-27a-M 14 miR-27a-I 1.6 ** H37Rv (MOl=5) * H37Rv (MOl=5) 12 1.4 ** NS ** 1.2 NC miR-27a-M NC miR-27a -l ) 10 4 1.0 0 13612013612 8 0 13612013 612(h) 0.8 6 LC3I 0.6 15 kDa LC3II CFU (×10 4 0.4 1.00 1.14 2.83 4.32 2.66 0.39 0.44 0.75 0.93 0.73 1.00 2.43 1.86 2.32 1.96 0.69 0.64 2.25 4.77 3.13 2

Bacteria survival (ratio) 0.2 GAPDH 0 0.0 35 kDa 2 h 24 h -I NC -M H37Rv (MOl=5) miR-27a miR-27a k H37Rv (MOl=5) hi 0 h 6 h H37Rv (MOl=5) 25 Autophagosome 0 h 6 h Autolysosome 20 WT 15

NC 10 –/– Puncta per cell 5

0 miR-27a -M NC 0 h NC 6 h

miR-27a-MmiR-27a-I 0 h 0 h miR-27a-MmiR-27a-I 6 h 6 h mRFP-GFP-LC3

miR-27a H37Rv (MOl=5) l 25 Autolysosome Autophagosome j WT miR-27a –/– 20

0 13612013612(h) 15 LC3 I 15kDa LC3 II 10 miR-27a-I 0.36 1.03 0.90 1.35 0.88 1.00 1.28 1.62 2.77 3.10

35kDa GAPDH Puncta per cell 5 mRFP-GFP-LC3 H37Rv (MOl=5) 0

0 h 6 h –/– –/– WT 0 h WT 6 h

miR-27a miR-27a dramatically decreased in miR-27a-/- macrophages (Supplemen- It has been reported that the intracellular survival of Mtb is tary Fig. 2c and d). LDH assay showed that there was no cyto- controlled by autophagy14. To examine whether miR-27a toxicity of miR-27a-/- macrophages induced by Mtb infection enhances Mtb survival through autophagy, wild type (WT) or (Supplementary Fig. 2e) and CFU assay showed that genetic miR-27a-/- macrophages, or macrophages transfected with deletion of miR-27a in peritoneal macrophages also reduced the miR-27a mimic or negative control (NC) were first pretreated intracellular survival of Mtb (Supplementary Fig. 2f and g). with either Bafilomycin A1 (Baf A1)34, a vacuolar-ATPase

NATURE COMMUNICATIONS | (2018) 9:4295 | DOI: 10.1038/s41467-018-06836-4 | www.nature.com/naturecommunications 3 ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-06836-4

Fig. 1 miR-27a promotes the intracellular survival of Mtb through autophagy. a–c 2D cluster analysis across miRNA probe (right) and subjects (bottom). Heatmap shows downregulated (green) and upregulated (red) miRNA from: (a) PBMCs of healthy donors (n = 3) and pulmonary TB patients (n = 3); (b) lung of the control or mice (n = 3) aerosol-infected with ~200 CFU Mtb H37Rv for 0 and 28 days; (c) murine primary peritoneal macrophages infected with H37Rv (MOI = 5) for indicated times. d–i macrophages pre-treated with NC, miR-27a-M (mimic), or miR-27a-I (inhibitor) were infected with Mtb for indicated times, and then subjected to (d, e) CFU detection; (f, g) Immunoblot (IB) of lysates; (h) confocal analysis of mRFP-GFP-LC3B spots, Bar, 5 μM. (i) Quantification of average number of LC3 puncta in each cell. (j–l)WTormiR-27a-/- macrophages infected with Mtb, and then subjected to (j) IB of lysates; (k) confocal analysis of mRFP-GFP-LC3B spots; Bar, 5 μM. l Quantification of average number of LC3 puncta in each cell. (d, e, i, l)**p < 0.01 by the unpaired t-test. Data are from three independent experiments with biological duplicates in each (d, e, i, l; mean ± s.e.m. of n = 3 duplicates) or representative of three independent experiments (f, g, h, j, k)

inhibitor that suppresses the autophagic flux at the Manipulation of miRNAs to promote the host innate defense is autophagosome–lysosome fusion stage, or AR-1235, an autophagy a promising strategy in the development of novel therapeutic inducer mainly targeting PDK-1, and were then infected with interventions against TB38,39. To investigate whether the Mtb. The altered effect of miR-27a mimic or miR-27a deficiency beneficial effects of miR-27a suppression could be translated into on the intracellular survival of Mtb was not found in a host-directed therapy (HDT), we used a miR-27a antagomir to macrophages pretreated with Baf A1 or AR-12 (Supplementary inhibit miR-27a in a mouse Mtb infection model and tested its Fig. 3a and b). Furthermore, WT and Atg5-/- Raw 264.7 therapeutic efficacy. First, the mice were infected with Mtb via the macrophages were transfected with miR-27a mimic or inhibitor, aerosol route for 14 days, and were then intraperitoneally injected and then infected with Mtb. Similarly, the altered effect of miR- with a NC antagomir or miR-27a antagomir solution at 5 mg/kg 27a mimic or inhibitor on the intracellular survival of Mtb was no for alternate days. At 28 days post-infection, qPCR results showed longer found in Atg5-/- macrophages (Supplementary Fig. 3c and that miR-27a antagomir treatment dramatically suppressed miR- d). These results suggested that miR-27a may promote the 27a expression in lung tissues and spleens of the mice, and intracellular survival of Mtb through modulating autophagy. increased the expression of Plk2 and Pink1, the two characterized It has been shown that conversion of soluble LC3-I to lipid miR-27a’s targets, in the lung tissues (Supplementary Fig. 5d and bound LC3-II is associated with the formation of autophago- e)40,41. Furthermore, the mice treated with miR-27a antagomir somes, and such conversion is a key marker of autophagy36. harbored a relatively lower bacterial load and displayed less Overexpression of miR-27a markedly reduced the amount of pathological impairment in their lungs (Fig. 2d, e, Supplementary LC3-II, whereas the miR-27a inhibitor enhanced the amount of Fig. 5f & g). Together, these results demonstrated a therapeutic LC3-II (Fig. 1f, g). To further examine the functional role of miR- effect of miR-27a antagomir against Mtb infection in the murine 27a in autophagy flux during Mtb infection, we used the mRFP- model. GFP-LC3 reporter construct to monitor LC3 aggregation in To further evaluate whether the therapeutic effects of miR-27a autophagosomes and autophagolysosomes:37 In pH-neutral antagomir treatment are mediated through miR-27a, we infected autophagosomes, mRFP-GFP-LC3 produces a yellow signal, the WT mice or miR-27a-/- mice with Mtb H37Rv strain for whereas in autophagolysosomes, a stronger red fluorescence 14 days, and then treated with NC antagomir or miR-27a signal is much more evident due to the loss of the pH-sensitive antagomir for another 14 days. Treatment with the miR-27a GFP signal. In macrophages infected with Mtb, miR-27a mimic antagomir lead to a reduction in bacterial load and less markedly inhibited the formation of autophagosomes, whereas pathological impairment in the lungs of WT mice, but not in the miR-27a inhibitor accelerated autophagosome maturation as miR-27a -/- mice (Figs. 2d, e, supplementary Fig. 5h). Similarly, determined by the number of LC3 puncta and the extent of the decreased bacterial burden caused by the miR-27a antagomir yellow-to-red conversion of the mRFP–GFP tandem-LC3 repor- in the lung of WT mice was not observed in the miR-27a-/- mice ter37 (Fig. 1h, i). Gene deletion of miR-27a in macrophages also infected with Mtb (Fig. 2f), suggesting that the anti-TB markedly enhanced the production of LC3-II and the therapeutic effect of the miR-27a antagomir may function autophagosome–autophagolysosome conversion (Fig. 1j–l). through inhibition of miR-27a. Furthermore, WT and miR-27a-/- macrophages were treated with chloroquine (CQ) and then infected with Mtb at MOI = 5, miR-27a directly targets Cacna2d3. Using base alignment western blot results and mRFP-GFP-LC3 reporter assay showed approach, a series of transcripts were found as potential targets of that deficiency of miR-27a increased autophagosome formation miR-27a, including Plk2, Ror1, St6galnc3, Trim23, Akirin and and LC3-II amount in the presence of CQ (Supplementary Cacna2d3(Cac3), and we chose Cac3 for the next-step study. Fig. 4a–c). Together, these data suggested that miR-27a may Cacna2d3, which encode a subunit of the voltage-dependent suppress the autophagic response to Mtb infection. 42,43 fi calcium (Cav) channel complex , was found to have a speci c binding site for miR-27a in its 3′-untranslated region (3′-UTR) miR-27a confers susceptibility to Mtb infection. To assess the (Fig. 3a). To confirm whether Cacna2d3 is a direct target of miR- functional role of miR-27a in the pathogenesis of TB in vivo, WT 27a, we generated luciferase reporter constructs by cloning either mice or miR-27a-/- mice were infected with the Mtb H37Rv strain the wildtype or a seed sequence-mutated (mut) 3′-UTR of Cac- for 28 days, and the bacterial burden and histopathological na2d3 into the 3′-UTR of a pGL-3M-Luc vector. The mutated 3′- impairment in the lung of the infected mice were analyzed. UTR of Cacna2d3 had four bases changed (GUACUGUA to Robust infiltration of neutrophil and lymphocyte were found in GGACGGGU), at the putative miR-27a-binding sites (Fig. 3a). the lung tissue of WT mice, but miR-27a-/- mice had less We co-transfected these vectors together with the miR-27a mimic pathological impairment and infiltration of neutrophil and lym- into HEK293T cells and analyzed the lysates 48 h later. Trans- phocyte in their lungs (Fig. 2a, supplementary Fig. 5a–c). More- fection with the miR-27a mimic markedly inhibited the luciferase over, compared to WT mice, miR-27a-/- mice infected with Mtb activity for the wild-type 3′-UTR of Cacna2d3, but showed no exhibited a decreased bacterial burden in their lungs as indicated significant repressive effect on the mutated 3′-UTR of Cacna2d3 by both the CFU assay and acid-fast staining (Fig. 2b, c). when compared to the control dsRNA (Fig. 3b). These results

4 NATURE COMMUNICATIONS | (2018) 9:4295 | DOI: 10.1038/s41467-018-06836-4 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-06836-4 ARTICLE

a PBS H37Rv WT miR-27a–/– WT miR-27a–/– Score: 0 Score: 0 Score: 4 Score: 1

b PBS H37Rv WT miR-27a–/– WT miR-27a–/–

cdH37Rv NC antagomir miR-27a antagomir WT miR-27a–/– Score: 4 Score: 2 8 ****

6 WT *** 4

Log10 (CFU) 2 Score: 3 Score: 2 –/–

0 1728 miR-27a Days post infection

efH37Rv WT miR-27a–/– NC antagomir miR-27a antagomir WT anti-miR-27a miR-27a–/– anti-miR-27a

8 ****

WT NS

6 **

4 NS –/– Log10 (CFU) 2

miR-27a 0 1728 Days post infection

Fig. 2 miR-27a controls susceptibility to Mtb. a–c WT or miR-27a-/- mice were infected with Mtb H37Rv for 28 days, and the lungs were subjected to (a) H&E staining. Scale bar, 100 μm; (b) acid-fast staining of bacteria. Scale bar, 20 μm. (c) CFU assay. d–f Mice were infected with Mtb for 28 days, and treated with NC antagomir or miR-27a antagomir alternate days, and the lungs were subjected to (d) H&E staining, Scale bar, 100 μm; (e) Acid-fast staining of bacteria. Scale bar, 20 μm. (f) CFU assay. *p < 0.05 and **p<0.01 by the Mann-Whitney U test (c, f). Data are representatives of three independent experiments with biological duplicates in each (c, f; mean ± s.e.m. of n = 6 duplicates) or representative of three independent experiments (a, b, d, e)

NATURE COMMUNICATIONS | (2018) 9:4295 | DOI: 10.1038/s41467-018-06836-4 | www.nature.com/naturecommunications 5 ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-06836-4

abcNC mir-27a-M 1.8 1.5 *** 1.6 NS ** 3′ 5′ miR-27a 1.4 1.2 ** 1.0 mRNA 5′ 3′ WT-Cac3 1.0 0.8 5′ 3′ Mut-Cac3 0.6 0.5

0.4 Cacna2d3 0.2 Relative expression of Relative luciferase activity 0.0 0.0 WT-UTR Mut-UTR 0 h 3 h 6 h 12 h H37Rv (MOI = 5)

defgNC miR-27a-M NC NC 2.0 14 miR-27a-I 130 kDa *** 1.8 miR-27a-M CACNA2D3 4 12 95 kDa 1.6 ** *** 1.00 0.42 1.4 10 3 1.2 GAPDH mRNA mRNA 8 *** 35 kDa 1.0 ** 2 6 mRNA level 0.8 h NC miR-27a-I 0.6 4 130 kDa Cacna2d3 1 Cacna2d3 0.4 2 CACNA2D3 Relative expression of Relative expression of 0.2 95 kDa Cacna2d3 0 0.0 0 1.00 1.53 Healthy TB patients 0 h 6 h 0 h 6 h 35 kDa GAPDH H37Rv (MOI = 5) H37Rv (MOI = 5)

ij kl WT miR-27a–/– WT miR-27a–/– 2.5 14 15 *** **** **** 12 *** 2.0 ** *** –/– 10 WT miR-27a 10 mRNA mRNA 1.5 8 1# 2# 3# 1# 2# 3# 130 kDa 2.0 ** CACNA2D3 1.0 95 kDa 1.5 miR-27a 5 GAPDH 0.5 1.0 35 kDa Cacna2d3 Cacna2d3 0.5 Relative expression of Relative expression of Relative expression of 0.0 0.0 0 0 h1 h3 h6 h 0 h 4 h 10 h 15 h WT miR-27a–/– PGN H37Rv (MOI = 5)

Fig. 3 miR-27a directly targets Cacna2d3. a Predicted binding sites of miR-27a and Cacna2d3 3′-UTR. b Luciferase activity of HEK293T cells transfected with either WT or the seed sequence-mutated (mut) 3′-UTR of Cacna2d3 plus scramble and miR-27a mimic. c–f Real-time PCR of the Cacna2d3 mRNA in: (c) macrophages infected with Mtb;(d) PBMC from healthy donors and TB patients; (e, f) macrophages pre-treated with miR-27a-M (e) or-I (f), and then infected with Mtb. g, h IB of the lysates from macrophages pre-treated with miR-27a-M or miR-27a–I, and then infected with Mtb. i Real-time PCR of Cacna2d3 in WT or miR-27a-/- macrophages infected with Mtb. j Real-time PCR of Cacna2d3 in WT and miR-27a-/- macrophages treated with PGN for indicated times. k Real-time PCR of Cacna2d3 in lung tissues from WT and miR-27a-/- mice. l IB of lysates from WT and miR-27a-/- macrophages (n = 3). NS, not significant (p > 0.05), *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001 by the unpaired t-test (b–f, i, j, k). Data are from three independent experiments with biological duplicates in each (b–f, i, j, k; mean ± s.e.m. of n = 3 duplicates) or representative of three independent experiments (g, h, l) suggest that miR-27a may suppress the expression of Cacna2d3 slightly increased in Mtb-infected miR-27a-/- macrophages, when by binding to the 3′-UTR of Cacna2d3 in a direct and sequence- compared to WT macrophages (Supplementary Fig. 7a–e), specific manner. suggesting that the induction of miR-27a may also suppressed The mRNA level of Cacna2d3 was much lower in murine expression of Plk2 and St6galnac3 during Mtb infection. primary macrophages infected with Mtb, and in PBMCs from active TB patients (Fig. 3c, d) compared to their representative controls. Pearson’s correlation analysis revealed that the expres- Cacna2d3 inhibits intracellular survival of Mtb. To dissect the sion of miR-27a was negatively correlated with the mRNA level of functional role of CACNA2D3 in the regulation of Mtb infection, Cacna2d3 in TB patients (Supplementary Fig. 6). To examine the we performed a CFU assay to analyze its effects on the intracel- effect of miR-27a on the expression of Cacna2d3 during Mtb lular survival of Mtb in macrophages. Transfection of macro- infection, murine primary macrophages were transfected with the phages with specific siRNAs for Cacna2d3 significantly increased miR-27a mimic or inhibitor and were then infected with Mtb. Mtb viability (Fig. 4a, Supplementary Fig. 8a & 8c). Furthermore, Treatment with the miR-27a mimic significantly inhibited the we introduced Cacna2d3 mutant mice which contain a piggyback mRNA and protein levels of Cacna2d3, whereas the miR-27a (PB) insertion in the intron region between exon 27 and inhibitor markedly enhanced both the mRNA and protein level of exon 28 and disrupted Cacna2d3 expression (Supplementary Cacna2d3 in Mtb-infected macrophages (Fig. 3e–h). Similarly, Fig. 8b) 44–46. Similarly, Suppression of Cacna2d3 by PB insertion genetic deletion of miR-27a increased the mRNA level of (Cac3PB/PB) also increased the intracellular survival of Mtb in Cacna2d3 in PGN-treated macrophages (Fig. 3i), as well as the primary peritoneal macrophages (Fig. 4b, Supplementary Fig. 8d). mRNA and protein levels of Cacna2d3 in Mtb-infected macro- Transmission electron microscopy (TEM) analysis revealed that phages, and lungs of mice infected with Mtb as compared to their relatively more bacteria are accumulated in Cac3PB/PB macro- representative controls (Fig. 3j–l). These results suggested that phages compared to WT cells (Fig. 4c, Supplementary Fig. 8e). induction of miR-27a suppress the expression of Cacna2d3 Cac3PB/PB macrophages also showed a reduced amount of LC3-II during Mtb infection. Furthermore, of the other candidate genes when infected with Mtb (Fig. 4d). Confocal microscopy analysis of miR-27a, only expression of Plk2 and St6galnac3 mRNAs were showed that inhibition of Cacna2d3 either by siRNA or PB

6 NATURE COMMUNICATIONS | (2018) 9:4295 | DOI: 10.1038/s41467-018-06836-4 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-06836-4 ARTICLE

H37Rv (MOI = 5) ab c dH37Rv (MOI = 5) 1.6 1.6 ** 0 h 24 h 1.4 ** 1.4 WT Cac3PB/PB 1.2 1.2 036 12 0 3 6 12 (h) 1.0 1.0 WT 15 kDa LC3 I 0.8 0.8 LC3 II 0.6 0.6 1.00 1.07 2.36 3.55 1.13 1.08 0.95 1.23 0.4 0.4 42 kDa Actin 0.2 0.2 Bacteria survival (ratio) Bacteria survival (ratio) 0.0 0.0 PB/PB siRNA: NC Cac3 WT Cac3m/m Cac3

efH37Rv (MOI = 5) gH37Rv (MOI = 5) 0 h 6 h 0 h 6 h

15 Autophagosome Autolysosome ** WT NC 10

5 Puncta per cell PB/PB 0 Cac3 Cac3

mRFP-GFP-LC3 siRNA 0 h siRNA 6 h mRFP-GFP-LC3 NC siRNA 0 h NC siRNA 6 h Cac3 Cac3

hijkWT WT BPD Cac3m/m Cac3m/m BPD *** Autophagosome 10 1.6 NS NS 15 Autolysosome WT 1.4 *** 2.5 8 Cac3PB/PB NS 1.2 NS ** ) 2.0 4 10 6 1.0 1.5 0.8 1.0 4 5 0.6 CFU (×10 0.5 0.4 Puncta per cell 2

Fluorescence (fold) 0.2 0 0 Bacteria survival (ratio) H37Rv 0 s 20 s 40 s 60 s 80 s 0 0.0 (MOI = 5) 0 h 6 h 2 h 24 h WT 0 h m/m WT 6 h m/m H37Rv (MOI = 5) m/m H37Rv (MOI = 5) BPD Cac3 Cac3 Cac3 WT BPDm/m WT DMSO Cac3

lmWT Cac3PB/PB BPD –+–+15 Autophagosome Autolysosome *** 0 h 10

5 NS Puncta per cell

0

H37Rv (MOI = 5) 6 h

BPD 0 h BPD 6 h WT BPD DMSO 0 hm/m 0 h WT BPD DMSO 6 hm/m 6 h WT DMSO 0m/m h WT DMSO 6m/m h Cac3 Cac3 mRFP-GFP-LC3 Cac3 Cac3

noH37Rv (MOI = 5) H37Rv (MOI = 5) WT Cac3PB/PB WT Cac3PB/PB 0 5 153060900 515306090(m) 0 0.5 1 3 6 120 0.5 136 12 (h) 55 kDa p-CaMKIIβ 130 kDa p-ULK1 (S555) 1.0 1.93 1.80 1.14 1.02 1.08 0.62 1.16 1.13 1.09 0.69 0.42 1.0 1.12 1.53 1.79 2.05 0.88 0.43 0.66 0.53 0.55 0.49 0.19 p-ULK1 (S757) 130 kDa 35 kDa GAPDH 1.0 1.03 1.00 1.04 1.62 1.45 0.96 0.93 1.43 1.89 2.32 2.85 42 kDa Actin suppression significantly reduced the accumulation of LC3 Mtb infection markedly increased the intracellular concentra- puncta and autophagy flux (Fig. 4e–h). Furthermore, WT and tion of Ca2+ flux. Treatment of Ca2+ channel blockers Cac3PB/PB macrophages were treated with CQ and then infected nilvadipine47 (Supplementary Fig. 9a) or amlodipine48 (Supple- with Mtb, western blot results showed that PB suppression of mentary Fig. 9d) markedly reduced the intracellular concentra- Cac3 also decreased LC3-II amount in the presence of CQ tion of Ca2+ in Mtb-infected macrophages. Inhibition of Ca2+ (Supplementary Fig. 8f). These results suggested that Cac3 may also reduced the amount of LC3-II (Supplementary Fig. 9b & 9e) promote autophagy and decrease the intracellular survival of Mtb. and the accumulation of LC3 puncta (Supplementary Fig. 9c

NATURE COMMUNICATIONS | (2018) 9:4295 | DOI: 10.1038/s41467-018-06836-4 | www.nature.com/naturecommunications 7 ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-06836-4

Fig. 4 Cacna2d3 regulates survival of Mtb via Ca2+-mediated autophagy. a, b Bacterial survival of Mtb in (a) primary macrophages treated with Cacna2d3 siRNA; (b) macrophages from WT and Cacna2d3 mutant (Cac3PB/PB) mice, after infection with Mtb (MOI = 5) for 24 h, compared to 2 h. c, d WT or Cac3PB/PB macrophages were infected with Mtb, and then subjected to (c) TEM analysis; red arrow indicated the bacteria. (d) IB of lysates. e–h Confocal detection of mRFP-GFP-LC3B spots in Cacna2d3-siRNA-treated macrophages (e, f)orCac3PB/PB macrophages, after infection with Mtb (MOI = 5) for 6 h (g, h); Bar, 5 μM. i Calcium assay in WT or Cac3PB/PB macrophages infected with Mtb (MOI = 5) for the duration of 1 min. j, k CFU assay (j) and intracellular survival of Mtb (k)inWTorCac3PB/PB macrophages pretreated with Bepridil (BPD) (10 μM), and then infected with Mtb (MOI = 5) for 2 h or 24 h. l, m Confocal detection of mRFP-GFP-LC3B spots in WT or Cac3PB/PB macrophages pretreated with DMSO or BPD infected with Mtb at MOI 5 for 6 h. Bar, 5 μM. n, o IB of lysates for phosphorylated-CaMKKβ (n) and phosphorylated-ULK1 (o) from WT or Cac3PB/PB macrophages infected with Mtb for indicated times. NS, not significant (p > 0.05), *p< 0.05 and ** p < 0.01 by the unpaired t-test (f, h) or Mann-Whitney U test (a, b, j). Data are from three independent experiments with biological duplicates in each (a, b, f, h, j, k, m; mean ± s.e.m. of n = 3 duplicates) or representative of three independent experiments (c–e, g, i, l, n, o)

& 9f), indicating an essential role of Ca2+ signaling in the CACNA2D3 promotes resistance to Mtb infection. To investi- regulation of autophagy during Mtb infection. CACNA2D3 is a gate the functional importance of CACNA2D3 in the regulation 2+ fl PB/PB subunit of the Cav channel that drives Ca in ux to activate of Mtb infection in vivo, WT or Cac3 mice were infected synaptic neurotransmitter release. Suppression of Cacna2d3 by with H37Rv at CFU = 200 for 28 days. The lung tissues of PB insertion markedly reduced the intracellular concentration of Cac3PB/PB mice showed more severe histology impairment and Ca2+ in Mtb-infected macrophages (Fig. 4i). To investigate more lymphocytes or neutrophil recruitment as compared to WT whether CACNA2D3 regulated the intracellular survival of Mtb mice (Fig. 5a, Supplementary Fig. 12). Furthermore, the bacterial through Ca2+ flux, we treated the WT or Cac3PB/PB macrophages titers in the lung tissues of Cac3PB/PB mice were much higher with the a membrane-permeable calcium channel blocker, than those of WT mice (Fig. 5b, c). Lastly, immunostaining of the Bepridil (BPD) 49, which could block the cytosolic calcium lung tissues of Cac3PB/PB mice showed less LC3 protein compared release, and then infected those cells with the Mtb H37Rv strain to WT mice (Fig. 5d). These results suggested that CACNA2D3 for 24 h. The inhibitory effect of CACNA2D3 on the Mtb viability may protect mice from Mtb infection by promoting autophagy. was not found in BPD-treated Cac3PB/PB macrophages (Fig. 4j, k). Similarly, treatment of BPD did not reduce the accumulation of LC3 puncta in the Cac3PB/PB macrophages as it did in WT miR-27a regulates Mtb infection via Cacna2d3 and Ca2+.To macrophages (Fig. 4l, m), suggesting that CACNA2D3 may determine whether miR-27a regulates the intracellular survival of + regulate the intracellular survival of Mtb via dampening Ca2 Mtb via targeting of Cacna2d3,WTorCac3PB/PB macrophages influx-mediated autophagy. were transfected with miR-27a mimic or inhibitor and were then + Regulation of autophagy by Ca2 is dependent on the infected with Mtb for 24 h. CFU assay and confocal microscopy + activation of CaMKK, a Ca2 /calmodulin-dependent serine/ analysis were performed to detect the intracellular bacterial load threonine kinase8. The phosphorylation level of CaMKKβ at and the autophagic responses, respectively. Transfection with the Thr286 was dramatically decreased in Mtb-infected Cac3PB/PB miR-27a mimic significantly increased the bacterial survival of macrophages compared to WT macrophages (Fig. 4n). When Mtb in WT macrophages, but not in Cac3PB/PB macrophages phosphorylated, CaMKKβ activates AMPK, which then phos- (Fig. 6a, Supplementary Fig. 13a). In contrast, the miR-27a phorylates ULK1 at Ser555 to initiate autophagy9. On the other inhibitor significantly reduced the bacterial survival of Mtb in WT hand, mTOR mediates the phosphorylation of ULK1 at Ser757 to macrophages, but not in Cac3PB/PB macrophages (Fig. 6b, Sup- inhibit autophagy by disassociating AMPK-ULK19. Suppression plementary Fig. 13b). Confocal microscopy analysis showed that of Cacna2d3 markedly reduced the phosphorylation of ULK1 on the formation of LC3 puncta in WT macrophages is inhibited by Ser555, but resulted in increased phosphorylation level on Ser757 the miR-27a mimic, but is enhanced by the miR-27a inhibitor. in Mtb-infected macrophages (Fig. 4o). It has also been shown However, no significant effect due to the miR-27a mimic or + that Ca2 signaling controls the activities of the phosphatase inhibitor on the formation of LC3 puncta in Cac3PB/PB macro- Calcineurin and of its substrate TFEB, a master transcriptional phages was observed (Fig. 6c, d). These results suggested that regulator of lysosomal biogenesis and autophagy11. Mtb infection miR-27a may inhibit autophagy to promote the intracellular markedly increased the nuclear translocation of TFEB in survival of Mtb by targeting Cacna2d3. Furthermore, RNAi of macrophages, but the neither the Cacna2d3 suppression Plk2 or St6galnac3, the other two candidate targets of miR-27a, nor the depletion of miR-27a had significant effect on the also increased the intracellular survival of Mtb in miR-27a-defi- activation of TFEB during Mtb infection (Supplementary Fig. 10). cient macrophages (Supplementary Fig. 13c). However, RNAi Together, these results suggest that CACNA2D3 may regulate of Pink1, another characterized miR-27a target in regulation + autophagy mainly through Ca2 -induced activation of CaMKKβ/ of mitophagy41, decreased intracellular survival of Mtb (Supple- ULK1. mentary Fig. 13d & e), suggesting that Pink1, may not be the + Intracellular Ca2 could be released from multiple organelles, downstream target of miR-27a in regulation of Mtb survival. including the ER, mitochondria, and lysosomes7. Immuno- Together, all these results indicated that miR-27a may simulta- staining analysis showed co-localization of CACNA2D3 with neously target multiple transcripts, resulting in potent cumulative the ER marker Calnexin, but not with the lysosome marker effects on Mtb viability. LAMP1 in Mtb-infected macrophages, suggesting that CAC- Xenophagy is a selective autophagy that delivers the Mtb into NA2D3 is mainly located on ER (Supplementary Fig. 11a, b). lysosome for degradation. We therefore examined whether miR- We therefore used the ER calcium chelator N,N,N’,N’-Tetrakis 27a inhibition or PB suppression of Cacna2d3 affect the co- (2-pyridylmethyl) ethylenediamine (TPEN) to examine its effect localization of Mtb H37Rv with p62, which tags bacteria for on autophagy flux during Mtb infection50. Treatment with TPEN sequestration within LC3-positive autophagosomes. Macrophages significantly reduced the number of LC3 puncta and autophago- transfected with miR-27a inhibitor contained more p62-positive lysosome formation (Supplementary Fig. 11c, d), suggesting an Mtb, when compared to those macrophages transfected with NC. + essential role for ER Ca2 in Mtb-induced autophagy. However, the increased p62-Mtb co-localization by miR-27a

8 NATURE COMMUNICATIONS | (2018) 9:4295 | DOI: 10.1038/s41467-018-06836-4 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-06836-4 ARTICLE

a PBS H37Rv

WT Cac3 PB/PB WT Cac3 PB/PB Score: 0 Score: 0 Score: 4 Score: 5

b PBS H37Rv WT Cac3 PB/PB WT Cac3 PB/PB

cdWT Cac3 PB/PB WT Cac3 PB/PB 10 **** 8 PBS ) 4 6 ** 4 CFU (×10 2

0 H37Rv 1 728 Days post infection

Fig. 5 Cacna2d3-mutant mice are more susceptible to Mtb infection. FVB and Cac3PB/PB mice were infected with Mtb or treated with PBS for 28 days, and the lungs were subjected to (a) H&E staining. Scale bar, 100 μm. b Acid-fast staining of bacteria. Scale bar, 20 μm. c CFU assay. d Immunohistochemistry of LC3B. Brown signal indicated LC3. Scale bar, 100 μm. **p < 0.01, ****p < 0.0001 by the Mann–Whitney U test (c). Data are representatives of three independent experiments. In c, data are representatives of three independent experiments with biological duplicates in each (c mean ± s.e.m. of n = 6 duplicates) inhibition was totally attenuated in Cac3PB/PB macrophages those of Mtb-infected Cac3PB/PB mice (Fig. 6l, m), suggesting that (Supplementary Fig. 14a, c). Likewise, macrophages transfected miR-27a may regulate Mtb infection by targeting Cacna2d3 with miR-27a inhibitor contains more LC3-positive Mtb when in vivo. compared to macrophages transfected with NC, while Cac3PB/PB macrophages contains less LC3-positive Mtb, even pre-treated with miR-27a inhibitor (Supplementary Fig. 14b, d). These results Discussion suggested that miR-27a may also suppress the xenophagy of Mtb Mtb is an intracellular pathogen that can evade lysosomal clearance via inhibiting Cac3. degradation and establish persistent infections. Although several We next examined whether miR-27a regulates the intracellular studies have shown that Mtb promotes intracellular survival by survival of Mtb through Ca2+ signaling. Transfection with the inhibiting autophagy through secretory proteins or the induction miR-27a mimic markedly inhibited the Mtb-induced Ca2+ influx, of miRNAs, but how Mtb manipulates Ca2+ signaling to regulate whereas the miR-27a inhibitor enhanced the Ca2+ influx (Fig. 6e). autophagy and promote its intracellular survival remains Similarly, in responses to Mtb infection, miR-27a-/- macrophages unknown. Here, we demonstrated that Mtb infection induces the showed enhanced Ca2+ influx compared to WT macrophages expression of miR-27a in Mtb-infected cells, animals and patients. (Fig. 6f). Pretreatment with the Ca2+ inhibitor BPD eliminated Induction of miR-27a direct targets and reduces the expression of the reduced survival of Mtb observed in miR-27a-/- macrophages the Ca2+ transporter subunit CACNA2D3, thus, downregulating compared to WT macrophages (Fig. 6g, h). The enhanced ER Ca2+ signaling to inhibit autophagy. These data not only formation of LC3 puncta was also not found in miR-27a-/- indicated miR-27a and CACNA2D3 as regulators of host macrophages pretreated with BPD (Fig. 6i, j). These results responses against Mtb infection, but also revealed a strategy for suggested that miR-27a may promote the survival the Mtb Mtb to manipulate the ER Ca2+ signaling to inhibit the host through downregulation of Ca2+ signaling. autophagy (Supplementary Fig. 15). Furthermore, treatment with To examine the relevance of miR-27a with Cacna2d3 in vivo, a miR-27a antagomir significantly ameliorated the pathogenesis WT or Cac3PB/PB mice were infected with Mtb for 14 days, and of Mtb infection, suggesting an intervention strategy for TB. were then treated with the NC antagomir or miR-27a antagomir Increasing studies have shown that miRNAs are involved in every 3 days for another 14 days. The lungs of infected mice were host responses to Mtb infection25,51. Our data demonstrated that then subjected to H&E staining. Treatment with the miR-27a miR-27a is not only highly expressed in Mtb-infected cells and antagomir alleviated the pathological impairments (Fig. 6k). animals, but is also much higher in active TB patients, suggesting Acid-fast staining and the CFU assay showed significantly lower that there is a clinical relevance of this microRNA with the bacterial titers in the lung of Mtb-infected WT mice, but not pathogenesis of TB. Induction of miR-27a inhibited the

NATURE COMMUNICATIONS | (2018) 9:4295 | DOI: 10.1038/s41467-018-06836-4 | www.nature.com/naturecommunications 9 ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-06836-4

a bcH37Rv (MOI = 5) dH37Rv (MOI = 5) 12 20 0 h 6 h 1.6 1.6 0 h 6 h 18 * 1.4 *** NS 1.4 *** NS 10 *** 16 1.2 1.2 8 Autophagosome 14 Autophagosome 1.0 1.0 Autolysosome 12 Autolysosome 0.8 0.8 6 10 8 0.6 0.6 NS NS 4 6

0.4 0.4 Puncta per cell Puncta per cell 2 4 0.2 0.2 2 Bacteria survival (ratio) Bacteria survival (ratio) 0.0 0.0 0 0

NC NC NC NC NC NC WT NC WT NC WT NC WT NC WT NC WT NC PB/PB PB/PB PB/PB PB/PB PB/PB miR-27-I PB/PB miR-27-I miR-27a-I miR-27a-M miR-27a-M miR-27a-M WT miR-27-IPB/PB WT miR-27-IPB/PB Cac3 WT miR-27a-ICac3 PB/PB Cac3 WT miR-27-M Cac3 WT miR-27-M Cac3 Cac3 WT miR-27a-MPB/PB PB/PB PB/PB Cac3 Cac3 Cac3 Cac3 Cac3 Cac3 e f g h NC WT miR-27a –/– DMSO WT DMSO WT DMSO –/– miR-27a-M –/– miR-27a DMSO miR-27a WT BPD miR-27a –/– BPD miR-27a-I WT BPD 5.0 14 NS 1.6 –/– 5.0 miR-27a BPD 12 NS 1.4 NS 4.0 4.0 *** 1.2 *** ) 10

4 * 1.0 3.0 3.0 8 0.8 6 2.0 2.0 0.6 CFU (×10 4 0.4 1.0

Fluorescence (fold) 1.0

Fluorescence (fold) 2 0.2 Bacteria survival (ratio) 0 0 0 0.0 0 s 20 s 40 s 60 s 80 s 0 s 20 s 40 s 60 s 80 s 2 h 24 h H37Rv (MOI = 5) H37Rv (MOI = 5) H37Rv (MOI = 5)

i DMSO BPD j

–/– –/– WT miR-27a WT miR-27a 25 Autolysosome Autophagosome ** 20

0 h 15 N.S 10 Puncta per cell 5

0 H37Rv (MOI = 5) 6 h

BPD 0 h BPD 6 h DMSO 0 h –/– DMSO 6 h–/– –/– WT BPD 0 h –/– WT BPD 6 h WT DMSO 0 h WT DMSOa 6 h

miR-27a miR-27a mRFP-GFP-LC3 miR-27a miR-27

klmWT Cac3PB/PB NC antagomir miR-27a antagomir NC antagomir miR-27a antagomir WT anti-miR-27a Score: 3 Score: 2 Cac3PB/PB anti-miR-27a 10 NS

WT WT 8 **** ) 4 6 NS Score: 4 Score: 4 ** 4 CFU (×10 PB/PB PB/PB 2 Cac3 Cac3 0 1728 Days post infection

Fig. 6 miR-27a regulates Mtb infection via Cacna2d3/Ca2+. a, b Intracellular survival of WT or Cac3PB/PB macrophages pretreated with (a) NC or miR-27a- M, (b) NC or miR-27a-I, and then infected with Mtb for 2 h and 24 h. c, d LC3 puncta from macrophages pretreated with NC and miR-27a-M (c)or miR-27a-I (d) then infected with Mtb. e, f Intracellular calcium fluorometry assay of (e) NC, miR-27a-M or miR-27a-I-treated, or (f)WTormiR-27a-/- macrophages, and then infected with Mtb for 1 min. g, h CFU assay (g) and intracellular survival of Mtb (h) of macrophages pretreated with DMSO or Bepridin (BPD), and then infected with Mtb for 2 and 24 h. i, j Assay of (i) mRFP-GFP-LC3B spots or (j) numbers of LC3 puncta in macrophages of WT and miR-27a-/- mice pretreated with DMSO or BPD, and then infected with Mtb. Bar, 5 μM. k, l FVB and Cac3PB/PB mice treated with NC antagomir or miR-27a antagomir twice a week for 1 month after Mtb infection for 15 days, and then subjected to (k) H&E staining of mice lung, Scale bar, 100 μm; (l) acid-fast staining of bacteria in the mice lung, Scale bar, 20 μm; m CFU in mice lung. NS, not significant (p > 0.05), *p<0.05 and ** p < 0.01 by the unpaired t-test (c, d, h, j) or Mann–Whitney U- test (a, b, g, m). Data are representatives from three independent experiments with biological duplicates in each (a, b, g, h, mean ± s.e.m. of n = 3 duplicates; c, d, j, m, mean ± s.e.m. of n = 6 duplicates) or representatives of three independent experiments (e, f, i, k, l)

10 NATURE COMMUNICATIONS | (2018) 9:4295 | DOI: 10.1038/s41467-018-06836-4 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-06836-4 ARTICLE autophagy to promote the intracellular survival of Mtb in mac- the activation of CaMKK/ULK1, but not the nuclear translocation rophages. Thus, the induction of miR-27a to inhibit the autop- of TFEB. Inhibition of ULK1 by RNAi is found to increase hagy appears to be a mechanism that could be used by Mtb as a intracellular survival of Mtb. Consistently, Horne et al. reported strategy to evade the host immune response. Furthermore, that ULK1 is genetically associated with Mtb, and ULK1-deficient treatment with a miRNA-27a antagomir protected mice from cells exhibited lower autophagy and higher Mtb replication59. Mtb infection as indicated by reduced pathological impairment in However, loss of ULK1 or other autophagy genes does not affect the lung and a lower bacterial load, suggesting that miRNA-27a bacterial burdens during Mtb infection in vivo19. The functional inhibitors could be repurposed for host-directed therapy of TB. role of autophagy genes in the restriction of Mtb replication is Though our data showed that TLR2 is essential for the complicated by the fact that Mtb encode effective inhibitors of induction of miRNA-27a during Mtb infection, but TLR2 only autophagy, as well as the additional roles of autophagy genes in confers marginal protection against mycobacterial infection52.It the regulation of inflammatory responses19,20,22. Indeed, ULK1 is possible that TLR2 activates multiple branches of downstream has been shown regulate the expression of TNF59. Further study signaling and stimulates both anti-bacterial effectors and pro- is required to clarify the functional role of CACNA2D3/ bacterial factors. The possible antagonism of these two arms of ULK1 signaling axis in the pathogenesis of TB. responses underscore the importance to specifically target factors Individual miRNAs may target multiple transcripts rather than involved in the immune evasion strategies employed by Mtb. one specific gene. Although several transcripts including Pink1 Other studies have reported that miR-27a targets various proteins and Plk2 were reported as miR-27a targets40,41,60, the relevance that participate in different cellular functions including tumor- between miR-27a and these genes need more validation. How- igenesis, hypoxia, and ischemia31,32. Moreover, whether miR-27a ever, our finding showed that Pink1 may not be required in the may exert its function in the pathogenesis of TB through targeting miR-27a-dependent regulation during Mtb infection. Also inhi- other proteins warrants further clarification. bition of St6galnac3 or Plk2 by RNAi attenuated the enhanced We found that miR-27a directly targets Cacna2d3,a bacterial killing of miR-27a-/- macrophages totally or partially, α2δ3 subunit of the Cav channel. Cacna2d3 was previously respectively. Possibly, miR-27a may simultaneously target mul- implicated to be involved in several tumors, including lung can- tiple transcripts including Cacna2d3, Plk2 and St6galnac3, cer, breast cancer, gastric cancer, and neuroblastoma53, however, resulting in potent cumulative effects on regulating autophagy a comprehensive study on the mechanism and extent its tumor and Mtb viability. suppressive function is still lacking. Our data demonstrated that Overall, our study have demonstrated that Mtb induce the Mtb downregulated the expression of Cacna2d3. Genetic inhibi- expression of miR-27a to downregulate Cacna2d3, which thereby tion of Cacna2d3 almost totally abrogated the intracellular Ca2+ inhibiting Ca2+-mediated autophagy, suggesting a strategy for influx, suggesting that the α2δ3-containing Cav channel as the Mtb to inhibit the host autophagy. Therefore, Mtb have evolved a main entry pathway for Ca2+ influx induced by Mtb infection. series of successful strategies for inhibition and evasion of host Interestingly, PB suppression of Cacna2d3 impaired autophago- autophagy. Indeed, the autophagy induction in response to Mtb some formation and was accompanied with higher intracellular infection is not so robust, which may shield the effect of deleting survival of Mtb. Furthermore, Cacna2d3-deficient mice had much host autophagy genes or blocking autophagy on Mtb growth. So, higher bacterial titers and were more susceptible to Mtb infection. the better way to examine the effect of autophagy on the Mtb Thus, our findings reveal a function for Cacna2d3 as an essential infection is to disarm the anti-autophagy strategies of Mtb. host autophagic factor for the Mtb control. Whether the autop- hagic effect of Cacna2d3 is associated with its anti-tumor function Methods remain to be explored. Clinical samples. All protocols were approved by the local ethics committee of Ca2+ is known to play a crucial role in TB pathogenesis. The + Tongji University School of Medicine (permit number: 2011-FK-03), and signed ability of Mtb to manipulate Ca2 signaling to prevent phago- informed consent was obtained from all subjects. Diagnosis of TB was based on some maturation and phagosome–lysosome fusion has been clinical presentation and radiological findings (such as an X-ray or computed fi extensively studied54–56, but the exact role of Ca2+ signaling in tomography (CT) scan) and was con rmed by a positive sputum culture. The test for the anti-human immunodeficiency virus (HIV) antibody was negative for all the regulation of autophagy during Mtb infection remains + TB patients. The average age was 40 years, and 65% of the patients were male. unknown. We found that a Ca2 blockade by specific inhibitors The healthy control individuals came from physical examination donors. The markedly reduced autophagosome formation and was accom- inclusion criteria were no history of previous TB or anti-mycobacterial treatments fi panied by a higher bacterial load of Mtb, suggesting an essential and no evidence of TB-related in ltration in chest X-rays. The average age was 35 years, and 60% of the controls were male. role of Ca2+ in autophagy and intracellular survival of Mtb. Intracellular Ca2+ is actively released to the cytosol from the ER, mitochondria, and lysosomes, and a more recent study revealed Cell lines. HEK293T cells and Raw264.7 cells were purchased from ATCC, and 2+ authenticated by the vendor. Atg5-/- Raw264.7 cell lines were gifted from Prof. Liu that lysosomal Ca release through MCOLN1 activates TFEB to ’ 11 Cuihua s Lab of Institute of Microbiology, Chinese Academy of Sciences. All cells regulate autophagy . We found that CACNA2D3 is mainly were mycoplasma-free with regular checks performed by a LookOut Mycoplasma 2+ located on the ER membrane, and that a blockade of ER Ca PCR Detection Kit (MP0035, Sigma-Aldrich). signaling markedly inhibits Mtb-induced autophagosome for- mation, suggesting a mechanism by which ER Ca2+ signaling is Mice. All animal study protocols were reviewed and approved by Tongji University linked to autophagy during Mtb infection. Of note, though being α δ School of Medicine review boards for animal studies. a regulatory 2 subunit of voltage gated calcium channels Cacna2d3PB/PB mice were generated by PB transposon and gifted by Prof. Wu required for the assembly of the channel on the membrane, the Xiaohui of Fudan University. Briefly, the PB transposon were inserted into the increase in the cellular abundance of CACNA2D3 only showed intron region between exon 27 and exon 28 of the Cacna2d3 gene of FVB mice. Cacna2d3 expression were disrupted due to the transcriptional termination signals modest (~2 fold) increasing effect on calcium signaling, most 42–44 PB/PB α δ within PB transposon . Cacna2d3 mice were backcross to FVB likely due to the loose association of 2 subunit with the chan- background more than 10 generations and maintained in heterozygotes. nel57,58 and a possible saturation of channel assembly on the The miR-27a-/- mice were generated through the CRISPR/Cas9 method33. membrane. Briefly, in vitro-translated Cas9 mRNA and gRNA were co-microinjected into the 2+ C57BL/6 zygotes. The three gRNA sequences used to generate the knockout mice is Activation of autophagic processes by Ca signaling is GAGAAGCCTATCATGACAAC, CTGTGAACACGACTTTGCTG, and dependent on the activation of CaMKK or TFEB. Our data + AGTGGCTAAGTTCCGCCCCC. Founders with frameshift mutations were demonstrated that Cacna2d3-mediated Ca2 signaling regulates screened with T7E1 assay and validated by DNA sequencing. Three F1 founders

NATURE COMMUNICATIONS | (2018) 9:4295 | DOI: 10.1038/s41467-018-06836-4 | www.nature.com/naturecommunications 11 ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-06836-4

-/- were got for the mir-27a mice, and one of them was chosen to backcross to the the miR-27a miRNA or mimic NCs (5′-UUCUCCGAACGUGUCACGUTT-3′) fi WT mice more than ve generations to maintain the strain. into HEK293T cells, and the luciferase assay was performed using the dual- C57BL/6 mice were purchased from Shanghai Laboratory Animal Center luciferase reporter assay system (Promega, Madison, USA). (Shanghai, China). All mice were bred in specific pathogen-free (SPF) conditions at the Laboratory Animal Center of Tongji University. Female mice, 6–8 weeks-of- age, were used. Aerosol infection of mice. All animal study protocols were reviewed and approved by Tongji University School of Medicine review boards for animal stu- dies. Infection studies were carried out using a murine respiratory infection Reagents and antibodies. The VGCC inhibitor Bepridil (sc-202974) was pur- model62–64. Mice were infected with ~200 CFU of Mtb H37Rv using a Glas-Col chased from Santa Cruz Biotechnology (Santa Cruz, USA).The vacuolar ATPase inhalation exposure system (Glas-Col Inc., Terre Haute, Ind.). At 28 days post inhibitor Bafilomycin A1 (S1413) and CQ (S4157) was purchased from Selleck infection, the mice were sacrificed and bacterial counts in the lungs were deter- Chem. (Shanghai, China). SB203580(559389), SP600125(420119), PD98059 mined by plating 10-fold serial dilutions of individual lung homogenateson Mid- (513000), and PDTC(548000) were purchased from Calbiochem (Merck, Darm- dlebrook 7H10 agar supplemented with 10% Middlebrook OADC enrichment. Mtb stadt, Germany) and were all used at a concentration of 10 μM. Adenovirus car- colonies were incubated at 37 °C and were counted after 21 days. rying mRFP-GFP-LC3B (C3011) or GFP-LC3B (C3006) were purchased from For miR-27a intervention, a miR-27a antagomir (5′- Beyotime Biotechnology (Shanghai, China). GCGGAACUUAGCCACUGUCAA-3′) and a NC antagomir were purchased from Rabbit anti-CACNA2D3 (ab102939, 1:1000 for WB and 1:100 for IF) and RIBOBIO (Guangzhou, China). Entranster TM-in vivo transfection reagent was mouse anti-Calnexin (ab31290, 1:100 for IF) were purchased from Abcam (Abcam, ordered from Engreen® (Beijing, China). The transfection protocol was modified Cambridge, USA); rabbit anti-LC3A/B (4108, 1:1000 for WB and 1:100 for IF amd based on manufacturer’s instructions. Briefly, reagent A was first prepared by IHC), rabbit anti-TFEB (37785, 1:1000 for WB and 1:100 for IF), rabbit anti-ATG5 dissolving 50 nmol of miR-27a antagomir or the NC antagomir in 375 μLof μ (12994, 1:1000), rabbit anti-p62 (8025, 1:1000), rabbit anti-Phospho-ULK1 autoclaved ddH2O, and then 375 L of a sterile 10% glucose solution was added (Ser555) (5869, 1:1000), rabbit anti-Phospho-ULK1 (Ser757) (14202, 1:1000), and and mixed well. Next, reagent B was prepared by adding 375 μL of sterile 10% rabbit anti-Phospho-CAMKII (Thr286) (12716, 1:1000) were purchased from Cell glucose solution to 375 μL of Entranster TM-in vivo transfection reagent, which Signaling Technology (CST, CA, USA). Monoclonal mouse anti-GAPDH was then mixed well. Finally, reagents A and B were mixed (1:1) to yield the (SAB2701826, 1:1000) was purchased from Sigma-Aldrich (St. Louis, MO, USA). working solution. For each injection, we used 300 μL of the working solution Monoclonal mouse anti-ACTINB (AC026, 1:5000) was purchased from Sigma- (~10 nmol of the miR-27a antagomir/NC antagomir). Aldrich (St. Louis, MO, USA).The HRP-conjugated anti-rabbit secondary antibody (W10804, 1:5000) was purchased from ThermoFisher Scientific (ThermoFisher Scientific, MA, USA). Hoechest33342 (H3570), Alexa Fluor488-conjugated anti- Transmission electron microscopy. Primary peritoneal macrophages from WT rabbit IgG (H + L) (A21206, 1:500 for IF) and Alexa Fluor 555-conjugated anti- FVB and Cacna2d3 mutant mice were infected with Mtb H37Rv strains at an MOI mouse IgG (H + L) (A32727, 1:500 for IF) were purchased from ThermoFisher of 5 for 6 h. Fixation and sectioning were performed using standard procedures by Scientific. the Microscopy Core at Shanghai Jiaotong University Medical School. Photographs were captured using HITACHI H-7650, and at least 20 images were acquired for each structure of interest; representative images are shown. PMBCs isolation. PBMCs were isolated from EDTA-treated whole blood through density gradient centrifugation by Ficoll separation (Ficoll-Paque plus; Amersham Mtb fl Biosciences). After two washes in PBS, the PBMCs were subjected to further Single cell suspensions of . Brie y, the H37Rv culture was dispersed by fi analyses. aspiration ve times each with a 23-gauge and then a 26-gauge needle, followed by an additional dispersion by aspiration three times through a 30-gauge needle. The dispersed bacteria were allowed to stand for 5 min to allow the clumps to settle Mouse macrophage isolation and infection. Peritoneal macrophages were iso- down. The upper half of the suspension was then used for the experiments. lated as described61. Briefly, mice were injected intraperitoneally (IP) with 2.0 mL Quantitation of bacteria was done by taking absorbance at 600 nm wavelength65 of 4% Brewer’s thioglycollate medium (B2551, Sigma-Aldrich). After 3 days, pri- (0.6 OD corresponds to ~100 × 106 bacteria). mary macrophages in the peritoneal lavage were collected from euthanized animals using 10 mL cold PBS. Subsequently, the cells were plated in 12-well plates at 106 cells/well in RPMI 1640 supplemented with 10% FBS and penicillin/streptomycin Histological analysis and acid-fast staining. Lung tissues from Mtb-infected mice were fixed in 4% phosphate-buffered formalin for 24 h and were then and were incubated at 37 °C in 5% CO2 for 2 h. The cultures were then washed fi fi three times with PBS to remove non-adherent cells, and the remaining adherent embedded in paraf n wax. The paraf n-embedded lungs were cut in serial sections with a thickness of 2–3 μm. Hematoxylin and eosin (H&E) staining were applied to monolayer cells were used as primary peritoneal macrophages. fi Mtb H37Rv (H37Rv) were grown to mid-log phase in Middle brook 7H9 detect the in ltration in the lungs. Acid-fast staining with the standard – medium (271310, Becton Dickinson, Cockeysville, MD) with 0.05% Tween-80 and Ziehl Neelsen method was employed for the determination of Mtb burden in the lungs. Five noncontiguous sections were examined for each lung, and at least five 10% oleic acid-albumin-dextrose-catalase (OADC) enrichment (211886, Becton fi Dickinson, Sparks, MD). Before infection, Mtb were suspended in complete animals were sacri ced for each group. The stained slides were visualized by light medium without antibiotics. Peritoneal macrophages were infected with H37Rv at microscopy. a MOI of 5. For immunohistochemistry, lung tissue from three PTB patients with pulmonary lobectomy and six Mtb-infected mice was investigated. Segments of lung tissues were fixed in 10% buffered formalin and were embedded in paraffin. Calcium ion flux measurement. To quantify the ion fluxes, fluorometry was The blocks were cut into 5 μm sections, and five noncontiguous sections were measured in Cellomics plate reader (Thermo Fisher Scientific, MA, USA). For processed for IHC staining with LC3B antibody at a 1:250 dilution. The sections calcium flux measurement, the Fluo-4 NW calcium assay kit (Invitrogen) was used. were then incubated with the Supersensitive 1-step polymer-HRP detection system Calcium fluxes were measured in 20,000 macrophages plated in a 96-well plate. (Biogenex). Immunostaining was visualized with 3,3-diaminobenzidine (DAB, Biogenex) substrate. After counterstaining in hematoxylin, the sections were mounted and examined. The studies on clinical samples were conducted in Immunofluorescence staining and confocal microscopy analysis. All images accordance with ethical guidelines of the Institutional Review Board of Tongji were captured on a Leica SP5 confocal microscope at ×63 magnification. For University. CACNA2D3 and TFEB staining, the primary peritoneal macrophages were fixed with 4% formaldehyde overnight, permeabilized with 0.1% Triton X-100 in PBS for 10 min, and blocked with 3% BSA in PBS for 30 min at RT. The cells were stained RNA interference and transfection. Mouse peritoneal macrophages were trans- with the indicated antibodies at a dilution of 1:200 in 3% BSA in PBS for 6 h at 4 °C fected with the small interference RNA, microRNA mimic or inhibitor with and were then incubated with Invitrogen Alexa Fluor 488 or 555 conjugated sec- INTERFER in the transfection reagent (AM4511, ThermoFisher). The siRNAs ondary antibodies at a dilution of 1:1000 for 2 h at RT. The nuclei were stained specifically targeting mouse Cacna2d3 and Tfeb were purchased from Santa Cruz. with Hoechst 33342. The siRNAs pool specifically targeting mouse Plk2, St6galnac3, and Pink1 were For mRFP-GFP-LC3B and GFP-LC3B analysis, primary peritoneal synthesized from Genepharma Co., Ltd (Shanghai, China) and the sequence are macrophages were pre-treated with adenovirus carrying mRFP-GFP-LC3B or GFP- listed in Supplementary Table 2. The miR-27a mimics and inhibitors were pur- LC3B for 48 h before further treatment. Then, the cells were imaged using confocal chased from RIBOBIO (Guangzhou, China). microscopy. LC3B spots were counted using the Cellomics ArrayScan platform. Q-PCR. For detection of Cacna2d3, total RNA from PBMCs and mice macro- Luciferase assay. A 0.5-kb region of the mouse Cacna2d3 3′-UTR containing the phages and lungs was extracted with TRIzol reagent according to the manu- predicted miRNA miR-27a-binding sites was cloned into psi-CHECK2 to obtain facturer’s instructions (Invitrogen). One microgram of total RNA was reverse the Cacna2d3 3′ UTR-luc construct. The Cacna2d3 3′-UTR (mut)-luc construct transcribed using the ReverTra Ace® qPCR RT Kit (Toyobo, FSQ-101) according to was obtained from the Cacna2d3 3′-UTR construct by mutating the com- the manufacturer’s instructions. A SYBR RT-PCR kit (Toyobo, QPK-212) was used plementary sequence of the miR-27a seed region (5′-ACUGUGA-3′ to 5′-AAGG for quantitative real-time PCR analysis. The relative mRNA expression of different ACA-3′). The constructs were co-transfected with synthesized dsRNAs mimicking genes was calculated by comparison with the control gene Gapdh (encoding

12 NATURE COMMUNICATIONS | (2018) 9:4295 | DOI: 10.1038/s41467-018-06836-4 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-06836-4 ARTICLE

GAPDH) using the 2-△△Ct method. The sequences of primers for the qPCR 16. Stanley, S. A. et al. Identification of host-targeted small molecules that restrict analysis are shown in Supplementary Table 1. intracellular Mycobacterium tuberculosis growth. PLoS Pathog. 10, e1003946 For the detection of miR-27a, miR-23a, and miR-24, PBMCs and mice (2014). macrophages and lungs were collected in TRIzol (Invitrogen) and were performed 17. Watson, R. O., Manzanillo, P. S. & Cox, J. S. Extracellular M. tuberculosis using the standard procedure. The miRNeasy Mini kit (QIAGEN) and TaqMan DNA targets bacteria for autophagy by activating the host DNA-sensing MicroRNA Reverse Transcription kit (Applied Biosystems) were used for total pathway. Cell 150, 803–815 (2012). RNA and cDNA preparation, respectively. Power SYBR Green master mix 18. Castillo, E. F. et al. Autophagy protects against active tuberculosis by (Applied Biosystems) was used for the qRT-PCR analysis on a 7900HT Fast Real- suppressing bacterial burden and inflammation. Proc. Natl Acad. Sci. USA Time PCR System (Applied Biosystems). Sno-RNA U18 was used as an internal 109, E3168–E3176 (2012). control. One-tailed t-tests were performed with GraphPad Prism 7.0, and a P- 19. Kimmey, J. M. et al. Unique role for ATG5 in neutrophil-mediated ≤ fi value 0.05 was considered signi cant. The sequences of primers were purchased immunopathology during M. tuberculosis infection. Nature 528, 565–569 from Tiangen Co., Ltd (Shanghai, China) and the Cat. no. are shown in (2015). Supplementary Table 1. 20. Zhang, L. et al. Effects of Mycobacterium tuberculosis ESAT-6/CFP-10 fusion protein on the autophagy function of mouse macrophages. DNA Cell Biol. 31, Western blot. Proteins were lysed from cells using RIPA buffer containing 10 mM 171–179 (2012). Tris–Cl, pH 8.0, 150 mM NaCl, 1% Triton X-100, 1% Na-deoxycholate, 1 mM 21. Saini, N. K. et al. Suppression of autophagy and antigen presentation by EDTA, 0.05% SDS and fresh 1 × proteinase inhibitor. The protein concentration Mycobacterium tuberculosis PE_PGRS47. Nat. Microbiol. 1, 16133 (2016). was determined via the Bradford method using the Bio-Rad protein assay before 22. Shui, W. et al. Organelle membrane proteomics reveals differential influence proteins were equally loaded and separated in polyacrylamide gels. The proteins of mycobacterial lipoglycans on macrophage phagosome maturation and were then transferred to a nitrocellulose filter membrane (Millipore) and were autophagosome accumulation. J. Proteome Res. 10, 339–348 (2011). incubated overnight with indicated primary antibodies. HRP-conjugated secondary 23. Bartel, D. P. MicroRNAs: genomics, biogenesis, mechanism, and function. Cell antibodies were then applied to the membrane, and the Western blot signal was 116, 281–297 (2004). detected using auto-radiographic film after incubation with ECL (GE Healthcare) 24. Xie, X. et al. Systematic discovery of regulatory motifs in human promoters fi or SuperSignal West Dura reagents (Thermo Scienti c). and 3’ UTRs by comparison of several mammals. Nature 434, 338–345 (2005). 25. Kim, J. K., Kim, T. S., Basu, J. & Jo, E. K. MicroRNA in innate immunity and Statistical analysis. The results are represented as the mean ± s.e.m., and statis- autophagy during mycobacterial infection. Cell Microbiol. 19, e12687 (2017). tical significance between groups was determined using an unpaired t-test or the 26. Ouimet, M. et al. Mycobacterium tuberculosis induces the miR-33 locus to Mann–Whitney U-test. GraphPad Prism software 7.0 was used for all analyses, and reprogram autophagy and host lipid metabolism. Nat. Immunol. 17, 677–686 a*p < 0.05 was considered statistically significant. (2016). 27. Kumar, R. et al. MicroRNA17-5p regulates autophagy in Mycobacterium tuberculosis-infected macrophages by targeting Mcl-1 and STAT3. Cell. Data availability Microbiol. 18, 679–691 (2016). Data have been deposited in the Gene Expression Omnibus under accession code 28. Kim, J. K. et al. MicroRNA-125a inhibits autophagy activation and GSE119494, GSE119495, GSE119496. Other data that support the findings of this antimicrobial responses during mycobacterial infection. J. Immunol. 194, study are available from the corresponding author upon request. 5355–5365 (2015). 29. Kim, J. K. et al. MIR144* inhibits antimicrobial responses against Received: 22 March 2018 Accepted: 27 September 2018 Mycobacterium tuberculosis in human monocytes and macrophages by targeting the autophagy protein DRAM2. Autophagy 13, 423–441 (2017). 30. Chen, W. et al. Emerging role of microRNA-27a in human malignant glioma cell survival via targeting of prohibitin. Mol. Med. Rep. 12, 1515–1523 (2015). 31. Towers, C. G. et al. The Six1 oncoprotein downregulates p53 via concomitant regulation of RPL26 and microRNA-27a-3p. Nat. Commun. 6, 10077 (2015). References 32. Zhao, W. et al. miR-27a-mediated anti-proliferative effects of metformin on 1. Kaufmann, S. H. et al. Progress in tuberculosis vaccine development and host- the breast cancer cell line MCF-7. Oncol. Rep. 36, 3691–3699 (2016). directed therapies--a state of the art review. Lancet Respir. Med. 2, 301–320 33. Wang, H. Y. et al. One-step generation of mice carrying mutations in multiple (2014). genes by CRISPR/Cas-mediated genome engineering. Cell 153, 910–918 2. World Health Organization. WHO Global Tuberculosis Report (2016). (2013). 3. Getahun, H., Matteelli, A., Chaisson, R. E. & Raviglione, M. Latent 34. Yoshimori, T. et al. Bafilomycin A1, a specific inhibitor of vacuolar-type H Mycobacterium tuberculosis infection. N. Engl. J. Med. 372, 2127–2135 (2015). (+)-ATPase, inhibits acidification and protein degradation in lysosomes of 4. Levine, B., Mizushima, N. & Virgin, H. W. Autophagy in immunity and cultured cells. J. Biol. Chem. 266, 17707–17712 (1991). inflammation. Nature 469, 323–335 (2011). 35. Chiu, H. C. et al. Eradication of intracellular Francisella tularensis in THP-1 5. Deretic, V., Saitoh, T. & Akira, S. Autophagy in infection, inflammation and human macrophages with a novel autophagy inducing agent. J. Biomed. Sci. immunity. Nat. Rev. Immunol. 13, 722–737 (2013). 16, 110 (2009). 6. Singh, R. & Cuervo, A. M. Autophagy in the cellular energetic balance. Cell. 36. Slobodkin, M. R. & Elazar, Z. The Atg8 family: multifunctional ubiquitin-like Metab. 13, 495–504 (2011). key regulators of autophagy. Essays Biochem. 55,51–64 (2013). 7. La Rovere, R. M., Roest, G., Bultynck, G. & Parys, J. B. Intracellular Ca(2+) 37. Kimura, S., Noda, T. & Yoshimori, T. Dissection of the autophagosome signaling and Ca(2+) microdomains in the control of cell survival, apoptosis maturation process by a novel reporter protein, tandem fluorescent-tagged and autophagy. Cell Calcium 60,74–87 (2016). LC3. Autophagy 3, 452–560 (2007). 8. Racioppi, L. & Means, A. R. Calcium/calmodulin-dependent protein kinase 38. Iannaccone, M., Dorhoi, A. & Kaufmann, S. H. Host-directed therapy of kinase 2: roles in signaling and pathophysiology. J. Biol. Chem. 287, tuberculosis: what is in it for microRNA? Expert Opin. Ther. Targets 18, 31658–31665 (2012). 491–494 (2014). 9. Kim, J., Kundu, M., Viollet, B. & Guan, K. L. AMPK and mTOR regulate 39. Man, D. K., Chow, M. Y., Casettari, L., Gonzalez-Juarrero, M. & Lam, J. K. autophagy through direct phosphorylation of Ulk1. Nat. Cell Biol. 13, 132–141 Potential and development of inhaled RNAi therapeutics for the treatment of (2011). pulmonary tuberculosis. Adv. Drug Deliv. Rev. 102,21–32 (2016). 10. Wong, P. M., Puente, C., Ganley, G. I. & Jiang, X. J. The ULK1 complex: 40. Tian, Y. et al. MicroRNA-27a promotes proliferation and suppresses apoptosis sensing nutrient signals for autophagy activation. Autophagy 9, 124–137 by targeting PLK2 in laryngeal carcinoma. BMC Cancer 14, 678 (2014). (2013). 41. Kim, J. et al. miR-27a and miR-27b regulate autophagic clearance of damaged 11. Medina, D. L. et al. Lysosomal calcium signalling regulates autophagy through mitochondria by targeting PTEN-induced putative kinase 1 (PINK1). Mol. calcineurin and TFEB. Nat. Cell Biol. 17, 288–299 (2015). Neurodegener. 11, 55 (2016). 12. Stanley, S. A. & Cox, J. S. Host-pathogen interactions during Mycobacterium 42. Catterall, W. A. Voltage-gated calcium channels. Cold Spring Harb. Perspect. tuberculosis infections. Curr. Top. Microbiol. Immunol. 374, 211–241 (2013). Biol. 3, a003947 (2011). 13. Cemma, M. & Brumell, J. H. Interactions of pathogenic bacteria with 43. Zamponi, G. W., Striessnig, J., Koschak, A. & Dolphin, A. C. The physiology, autophagy systems. Curr. Biol. 22, R540–R545 (2012). pathology, and pharmacology of voltage-gated calcium channels and their 14. Gutierrez, M. G. et al. Autophagy is a defense mechanism inhibiting BCG and future therapeutic potential. Pharmacol. Rev. 67, 821–870 (2015). Mycobacterium tuberculosis survival in infected macrophages. Cell 119, 44. Ding, S. et al. Efficient transposition of the piggyBac (PB) transposon in 753–766 (2004). mammalian cells and mice. Cell 122, 473–483 (2005). 15. Sakowski, E. T. et al. Ubiquilin 1 promotes IFN-γ-induced xenophagy of 45. Cui, J. et al. Disruption of GPR45 causes reduced hypothalamic POMC Mycobacterium tuberculosis. PLoS Pathog. 11, e1005076 (2015). expression and obesity. J. Clin. Invest. 126, 3192–3206 (2016).

NATURE COMMUNICATIONS | (2018) 9:4295 | DOI: 10.1038/s41467-018-06836-4 | www.nature.com/naturecommunications 13 ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-06836-4

46. Li, G. et al. The histone methyltransferase Ash1l is required for epidermal 65. Karim, A. F. et al. Express path analysis identifies a tyrosine kinase Src-centric homeostasis in mice. Sci. Rep. 7, 45401 (2017). network regulating divergent host responses to Mycobacterium tuberculosis 47. Yamazaki, H. et al. Preservation of retinal morphology and functions in royal infection. J. Biol. Chem. 286, 40307–40319 (2011). college surgeons rat by nilvadipine, a Ca(2+) antagonist. Invest. Ophthalmol. Vis. Sci. 43, 919–926 (2002). 48. Nimmrich, V. & Eckert, A. Calcium channel blockers and dementia. Acknowledgements Br. J. Pharmacol. 169, 1203–1210 (2013). We thank Prof. Xu Tian and Prof. Wu Xiaohui for the Cacna2d3 mutant mice. We thank 49. Gill, A., Flaim, S. F., Damiano, B. P., Sit, S. P. & Brannan, M. D. Pharmacology Prof. Liu Cuihua for the Atg5-/- Raw264.7 cells. We thank Feifan Xu (Nantong Sixth of bepridil. Am. J. Cardiol. 69, 11D–16D (1992). people’s Hospital, China) for samples collection. We acknowledge Dr. Xinchun Chen 50. Garrity, A. G. et al. The endoplasmic reticulum, not the pH gradient, drives (Shenzhen Third people’s Hospital) and members of the Xinchun Chen’s laboratory for calcium refilling of lysosomes. eLife 5, e15887 (2016). expert technical assistance of mice infection. We thank Dr. Gucheng Zeng (Sun Yat-Sen 51. Yang, T. & Ge, B. miRNAs in immune responses to Mycobacterium University) and members of his laboratory for expert technical assistance of mice tuberculosis infection. Cancer Lett. 431,22–30 (2018). infection. We thank members of the Shanghai Key Lab of Tuberculosis and Baoxue Ge’s 52. Feng, C. G. et al. Mice lacking myeloid differentiation factor 88 display laboratory for helpful discussions. This work was supported by National Natural Science profound defects in host resistance and immune responses to Mycobacterium Foundation of China (Nos. 81300003, 81501373, 91542111, and 81370108). avium infection not exhibited by Toll-like receptor 2 (TLR2)- and TLR4- deficient animals. J. Immunol. 171, 4758–4764 (2003). Author contributions 53. Palmieri, C. et al. Methylation of the calcium channel regulatory Subunit F.L., J.X.C., and B.X.G. designed experiments. J.X.C., F.L., P.W., H.H.L. Y.L.Z., H.P.L., CACNA2D3 predicts site-specific relapse in oestrogen receptor-positive Z.H.L., and R.J.Z. performed experiments and analyzed data. P.W. facilitated clinical primary breast carcinomas. Br. J. Cancer 107, 375–381 (2012). sample collection and transport. X.C.H. and W.J. prepared the bacteria H37Rv. R.J.Z., 54. Malik, Z. A., Denning, G. M. & Kusner, D. Inhibition of Ca2+ signaling by H.Z.L., F.W., and J.X.C. contributed to mouse experiments. J.X.C., F.L., Y.L.Z., H.P.L., Mycobacterium tuberculosis is associated with reduced phagosome-lysosome H.H.L., and B.X.G. prepared the figures. F.L., J.X.C., and B.X.G. wrote the manuscript. fusion and increased survival within human macrophages. J. Exp. Med. 191, 287–302 (2000). 55. Jayachandran, R. et al. Survival of mycobacteria in macrophages is Additional information mediated by coronin 1-dependent activation of calcineurin. Cell 130,37–50 Supplementary Information accompanies this paper at https://doi.org/10.1038/s41467- (2007). 018-06836-4. 56. Sharma, S. & Meena, L. S. Potential of Ca2+ in Mycobacterium tuberculosis H37Rv pathogenesis and survival. Appl. Biochem. Biotechnol. 181, 762–771 Competing interests: The authors declare no competing interests. (2017). 57. Cassidy, J. S., Ferron, L., Kadurin, I., Pratt, W. S. & Dolphin, A. C. Functional Reprints and permission information is available online at http://npg.nature.com/ exofacially tagged N-type calcium channels elucidate the interaction with reprintsandpermissions/ auxiliary α2δ-1 subunits. Proc. Natl Acad. Sci. USA 111, 8979–8984 (2014). 58. Gerald,W.,Zamponi,J.S.,Alexandra,K.&Annette,C.D.Thephysiology, Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in pathology, and pharmacology of voltage-gated calcium channels and fi their future therapeutic potential. Pharmacol. Rev. 67,821–870 published maps and institutional af liations. (2015). 59. Horne, D. J. et al. Human ULK1 variation and susceptibility to Mycobacterium tuberculosis infection. J. Infect. Dis. 214, 1260–1267 (2016). Open Access 60. Zhang, M., Sun, W., Zhou, M. & Tang, Y. MicroRNA-27a regulates hepatic This article is licensed under a Creative Commons lipid metabolism and alleviates NAFLD via repressing FAS and SCD1. Sci. Attribution 4.0 International License, which permits use, sharing, Rep. 7, 14493 (2017). adaptation, distribution and reproduction in any medium or format, as long as you give 61. Kong, L. et al. An essential role for RIG-I in toll-like receptor-stimulated appropriate credit to the original author(s) and the source, provide a link to the Creative phagocytosis. Cell Host Microbe 6, 150–161 (2008). Commons license, and indicate if changes were made. The images or other third party ’ 62. Kelly, B. P., Furney, S. K., Jessen, M. T. & Orme, I. M. Low-dose aerosol material in this article are included in the article s Creative Commons license, unless infection model for testing drugs for efficacy against Mycobacterium indicated otherwise in a credit line to the material. If material is not included in the ’ tuberculosis. Antimicrob. Agents Chemother. 40, 2809–2812 (1996). article s Creative Commons license and your intended use is not permitted by statutory 63. Zheng, R. J. et al. Notch4 negatively regulates the inflammatory response to regulation or exceeds the permitted use, you will need to obtain permission directly from Mycobacterium tuberculosis infection by inhibiting TAK1 activation. J. Infect. the copyright holder. To view a copy of this license, visit http://creativecommons.org/ Dis. 218, 312–323 (2018). licenses/by/4.0/. 64. Yang, H. et al. Lysine acetylation of DosR regulates the hypoxia response of Mycobacterium tuberculosis. Emerg. Microbes Infect. 7,34 (2018). © The Author(s) 2018

14 NATURE COMMUNICATIONS | (2018) 9:4295 | DOI: 10.1038/s41467-018-06836-4 | www.nature.com/naturecommunications